Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

chore: bump cuda base version to a more recent one #248

Merged
merged 1 commit into from
Oct 2, 2024
+3 −2
Merged

chore: bump cuda base version to a more recent one #248

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions changes/248.changed
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1 @@
cuda base docker image is now `nvidia/cuda:12.6.1-runtime-ubuntu24.04`
2 changes: 1 addition & 1 deletion substrafl/remote/register/register.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -37,7 +37,7 @@
"""

_GPU_BASE_IMAGE = """
FROM nvidia/cuda:11.8.0-runtime-ubuntu22.04
FROM nvidia/cuda:12.6.1-runtime-ubuntu24.04
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

11.8.0 -> 12.6.1 seems like a huge bump - are we sure there was not a reason for using 11.8.0?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

11.8 is not supported in Ubuntu 24.04.
Having to use Ubuntu 22.04 is forcing us to either install a vulnerable version of libpixman-1-dev (a dependency of openslides-tool) or to install libpixman-1-dev from source (I must admit, I haven't investigated much further this option).
Current versions of pytorch are still supporting CUDA 11.8, but it's the oldest version supported.

I can try with cuda 12.5.1, which is the oldest CUDA supported by Ubuntu 24.04 https://gitlab.com/nvidia/container-images/cuda/blob/master/doc/supported-tags.md

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Tested on dev ✅


# update image & install Python
ENV DEBIAN_FRONTEND=noninteractive
Expand Down
2 changes: 1 addition & 1 deletion tests/remote/register/test_register.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -56,7 +56,7 @@ def test_get_base_docker_image_cpu():

def test_get_base_docker_image_gpu():
expected_dockerfile = """
FROM nvidia/cuda:11.8.0-runtime-ubuntu22.04
FROM nvidia/cuda:12.6.1-runtime-ubuntu24.04

# update image & install Python
ENV DEBIAN_FRONTEND=noninteractive
Expand Down
Loading